Rod mediated dark adaptation, a functional test for early and intermediate AMD outcomes

KEYWORDS:

• Age-related macular degeneration
• fundus grading
• vision testing
• drusen
• fovea

1. Introduction

Approved strategies for reducing the burden of age-related macular degeneration (AMD) focus on preventing or stabilizing neovascularization and slowing the progression of preexisting atrophy. Yet the vast majority of people with AMD have earlier disease. Currently there are neither proven means to arrest the progression of early and intermediate AMD nor have endpoints suitable for interventions in early disease phases been identified. Based on our model of AMD pathophysiology, outlined below, we believe that preventing advanced AMD by targeting early disease is achievable. This editorial proposes that rod-mediated dark adaptation (RMDA) is a promising endpoint due to its biologic validity based on multidisciplinary studies of aging, AMD’s strongest risk factor.

2. Body

What is dark adaptation? It is a psychophysical test that assesses how sensitivity to light recovers over time following a very bright light presented to the retina, as reported by an observer (Figure 1). An initial short segment represents the recovery of cone photoreceptors, ≤5 min long in most normally sighted adults. A second longer segment represents the recovery of rod photoreceptors. RMDA reflects the rate of rhodopsin regeneration in rods, involving the classic visual cycle in the retinal pigment epithelium (RPE). Recovery speed is rate-limited by retinoid concentration. This concentration, in turn, is a manifestation of the steps of eliminating products of light absorption from photoreceptor outer segments, recycling of released retinoid to its original form (11-cis-retinal), and regenerating the visual photopigment opsin [3].

Figure 1. Dark adaptation in aging and AMD.

Representative plots from two older adults in normal macular health and four with early AMD, depicting the time-course of light sensitivity recovery for a target at 5° following a photo-bleach (of 83%) [1]. The early segment of cone-mediated sensitivity recovery (up to 5 min) is followed by rod-mediated sensitivity recovery. Arrows mark for each eye the rod-intercept time (RIT), a metric that signifies the speed of dark adaptation. RIT represents the time to restore 3 log units of sensitivity. Higher values for RIT indicate slower recovery. Note the dramatically slowed recovery in early AMD eyes compared to those in normal macular health. Eye AMD4 did not recover sensitivity during the 20-min protocol, and therefore, an X-axis intercept cannot be shown. Recently we showed in a large sample that 14 persons with intermediate AMD (15.7%), 4 persons with early AMD (3.1%), and 1 person with a normal macula (0.4%) did not reach RIT within 45 min [2].

In the 1990s, retinal neuroscience and drusen biology converged to highlight RMDA as a functional indicator of macular disease progression. In 1993 Alan Bird’s group showed that RMDA measured at 3° (~0.86 mm) from fixation markedly slowed in patients with early and intermediate AMD [4]. This investigation was prompted by their 1990 histochemistry study showing a prominent deposition of lipid in Bruch’s membrane in donor eyes aged 1–95 years [5]. They postulated a localized shortage of retinoids at the photoreceptors due to impaired transport from circulation across this hydrophobic barrier. Concurrently, the first two-dimensional digital maps of photoreceptor density in healthy young human eyes by Curcio et al established that the 6-mm-diameter central retina was rod-dominated (rod: cone ratio, 9:1). Rod densities were highest in an elliptical ring at 3–5 mm eccentricity at the arcades and surrounding the optic nerve head. In eyes >60 years, rod photoreceptors at 0.5–3 mm from the foveal center, on the inner slope of the rod ring, declined 30% while cones were relatively preserved. In advanced AMD eyes, the last remaining photoreceptors were cones [6,7]. The impaired barrier model was further supported by Jacobson, Cideciyan, Pugh, and Lamb in studying clinical conditions demonstrably involving localized deficiency of retinoids [8].

By 2000, Owsley et al classified AMD stages with a color fundus photography system (as opposed to clinical examination) and found that RMDA slows with each decade during adulthood [9]. Bird’s initial findings were confirmed and extended by demonstrations that RMDA was substantially delayed in AMD eyes relative to older adults in normal macular health [10,11]. Owsley et al also noted that persons with early AMD complained about night vision and tasks performed under dim illumination [12]. Several laboratories around the world independently confirmed and fortified these findings [13–15].

To determine what functional test best discriminates between aging and early AMD, an observational study must include many tests, with a substantial cohort of patients. ALSTAR2 (Alabama Study on Early Age-related Macular Degeneration 2, NCT04112667) is a prospective observational study on 532 persons aged ≥60 years with early and intermediate AMD and those in normal macular health. Tests at baseline and a three-year follow-up visit include photopic and mesopic acuity, photopic and mesopic contrast sensitivity, photopic, mesopic and scotopic microperimetry, and RMDA. Cross-sectional data from the baseline visit of ALSTAR2 clearly showed that the visual function that changed the most from normal aging to intermediate AMD was RMDA. Other visual functions displayed a fraction of the range displayed by RMDA [15,16]. RMDA significantly differed between early and intermediate AMD, and between normal aging and early AMD; no other visual function did the latter [16]. Longitudinal analyses now in progress will compare decreases in RMDA and other visual functions over 3 years and examine their associations with retinal structural changes.

The strong showing for RMDA in this direct comparison was predicted from the underlying hypothesis of ALSTAR2, a Center-Surround model of deposit-driven AMD [2] (Figure 2). This model incorporates steep gradients of photoreceptor cell density in central retina and the clustering of high-risk soft drusen material under the fovea, as learned from tissue-level studies and epidemiology [17]. Drusen are located between the RPE basal lamina and inner collagenous layer of Bruch’s membrane. It is proposed that the foveal centration reflects the high metabolic demand of foveal cones plus transfer of xanthophyll macular pigments (Figure 2). The latter are delivered by plasma lipoproteins, taken up by RPE for transfer to the foveal retina, with unneeded lipids released basally as lipoproteins that are trapped by aging Bruch’s and choriocapillaris endothelium. Figure 2 shows how spatially concentric opposing effects can produce a narrow central peak of cone resilience and a broad surround of rod vulnerability in aging and early AMD. RMDA is tested at 5°, where rod loss in aging is maximal. The 5° location is also on the rim of the high-risk drusen area.

Figure 2. Center-surround model of cone resilience and rod vulnerability in aging and AMD.

The top row shows en face views of central retina with the 6-mm-diameter ETDRS grid superimposed (central subfield is 1 mm diameter). The effect of AMD pathology on visual function is modeled as two opposing mechanisms with different spatial extents, both aligned on the fovea. A spatially narrow mechanism is the beneficial effect on foveal cones represented by highly concentrated xanthophyll carotenoid pigment (orange). A spatially broad mechanism is the harmful effect of high-risk soft drusen material (gray), within the central 3 mm diameter (inner ring ETDRS) as supported by histopathology and epidemiology. It is wider than just the central subfield because xanthophyll also localizes to inner retinal layers, all of which are believed to be supplied from the choroid across the RPE. In the bottom row, these effects are plotted on one vertical axis, where help is positive, and harm is negative. The dashed line represents the choriocapillaris. Together, help and harm make a narrow center of foveal cone resilience atop a broad surround of para- and perifoveal rod vulnerability (right column). Cones in the foveal center sit above soft drusen and precursor lipoproteins in aging Bruch’s membrane and are affected by this deposition. They are also protected because Muller glia act as a secondary source aiding metabolism of cones. The rods sitting above the rim of the high-risk area are those that exhibit the most severely delayed RMDA. The resemblance of the composite map (upper right) to maps of photoreceptor loss in human retinal aging [6] is striking.

Precisely placed visual stimuli probe not only different photoreceptor populations but also differential risks for distinct progression sequences. Different forms of neovascularization (type 1 vs type 3) and atrophy are preceded by accumulation and expansion of extracellular deposits on either side of the RPE. In contrast to drusen, subretinal drusenoid deposits (SDD; also called reticular pseudodrusen) are located between photoreceptors and the RPE and first appear superior to the fovea in the ring of high rod density (Figure 3). Due to the differential cholesterol composition of drusen and SDD, lipid transfer and cycling between photoreceptors and support cells are candidate dysregulated pathways [18]. Due to the topographic relationship of drusen to the fovea and SDD to the rod ring, these pathways may represent constitutive functions specific to each photoreceptor class. Of note, eyes with SDD were shown by Cukras and associates to have markedly worse RMDA than AMD eyes lacking SDD [14].

Figure 3. RMDA is affected near the fovea in eyes with subretinal drusenoid deposits (SDD).

Drusen are seen best by color fundus photography (left) and SDD by near-infrared reflectance (right). RMDA tested at 5° (bottom green spot) where rods are sparse slows more than at 12° (top green spot), where rods are numerous and SDD begins.

As we and others found that RMDA slowed more near the fovea than further away (up to 8° eccentric), ALSTAR2 explored regional differences to optimize test conditions (Figure 3). We tested 5° next to the foveal cones, where rods are relatively sparse (rod:cone ratio ~ 4:1) and disproportionately lost in aging, and at 12° where rods are numerous (rod:cone ratio 10–13:1) and SDD begins. We found that measuring RMDA at 5° differentiates much better between normal aging and intermediate AMD than measuring at 12°, even in eyes with SDD. Further, this difference increases with AMD severity through intermediate AMD [2]. Interestingly, MACUSTAR, an ongoing European multicenter study (NCT03349801), is seeking biomarkers for progression from intermediate AMD to geographic atrophy and is testing RMDA at 12° only. The 12° location was chosen in combination with a 76% bleach to separate diagnostic groups while potentially reducing the recovery time to under 20 min, compared to longer recovery times in the 5° location (up to 45 min, Figure 1). No specific model of pathophysiology to support this choice was mentioned.

3. Expert opinion

An international consensus group [19] has advocated for enhanced detection and diagnosis, among other measures, to enable early intervention in AMD. These steps are essential for reducing the treatment burden of advanced disease, which now also includes geographic atrophy with the recent approvals of two complement inhibitors. To implement RMDA as an outcome in AMD trials, several tasks must be accomplished.

1. The natural history of RMDA in early and intermediate AMD eyes must be characterized longitudinally to interpret the impact of any intervention directed toward these stages.

2. Training in administration of the RMDA test must be standardized, and procedures harmonized across sites for maximal comparability. Benchmarks for successful RMDA testing should be established by large sample studies from highly experienced investigators.

3. RMDA testing takes longer (Figure 1) than testing acuity and contrast sensitivity with letter charts. A shorter RMDA test sufficient for trials might be achieved with a lower intensity bleach and machine learning to optimize the procedure. Testing at greater eccentricities, as did MACUSTAR [2], will shorten the time for individual patients. It could also result in longer and more expensive trials due to the reduced effect size, and thus statistical power.

4. To be considered for a trial that meets regulatory requirements, RMDA should be validated in reference to the everyday experiences of patients.

5. In the near term, performance on RMDA can be used to benchmark candidate biomarkers and imaging technologies. Imaging tests are typically much quicker to administer than psychophysical tests. For example, in the ALSTAR2 baseline, greater choriocapillaris flow signal deficit measured with optical coherence tomography angiography correlates well with delayed RMDA, thus supporting an overall model of transport impairment that underlies both visual dysfunction and specific AMD pathology [20].

Declaration of interest

CA Curcio receives research funds from Genentech/Hoffman LaRoche and Regeneron and consults for Apellis, Astellas, Boehringer Ingelheim, Character Biosciences, Osanni, and Annexon (outside this project). C. Owsley is an inventor of the method and apparatus for the detection of impaired dark adaptation used in our research. She consults for Johnson & Johnson Vision (outside this project). D. Kar is currently an employee of Apellis Pharmaceuticals.

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

R01EY029595 (CO, CAC) and R01EY027948 (CAC); P30EY03039 (CO); Dorsett Davis Discovery Fund (CO), and Alfreda J. Schueler Trust (CO); Unrestricted funds to the Department of Ophthalmology and Visual Sciences (UAB) from Research to Prevent Blindness, Inc., and EyeSight Foundation of Alabama.